A major emphasis of this project has been to define the relationship of exogenous hormones to subsequent cancer risk. Recent analyses have assesed the relationships of menopausal hormones to gynecologic and breast cancer risk using data from our large, prospective cohort studies. Using data from the NIH-AARP Diet and Health Cohort Study, we have clarified some unresolved issues regarding ovarian cancer risk in women who use menopausal hormone therapy. Increased ovarian cancer risks among women who used unopposed estrogen therapy for 10 or more years provide further evidence to support the hypothesis that increased estrogen levels after menopause can influence the development of ovarian cancer. In addition, this study provided some of the first strong evidence that specifically links estrogen plus progestin use to increased ovarian cancer risk in women with intact uteri. These results reiterate the value of long-term follow-up of existing cohorts to elucidate increased risks of rare outcomes, such as ovarian cancer, among women who exposed to menopausal hormone therapy. This large cohort study has also been used to assess relationships of menopausal hormones to the risk of endometrial cancers. In contrast to some previous suggestions that estrogen plus progestin therapy might protect against this cancer, we found no evidence for such protection. In fact, we found evidence that women who used progestins sequentially, particularly for long periods of time, might be at an increased risk of developing endometrial cancer compared to non-users. Further, within this same study we examined risks of breast cancer related to both estrogens alone and with combined estrogen-progestin therapy. This investigation found elevated risks for both types of preparations, although combined therapy was more strongly related. Hormone effects were stronger among thin women, but combined therapy continued to be a risk factor even among heavy women. Finally, data were used to assess whether hormones had differential relationships on different types of tumors. The strongest effects were seen for estrogen receptor positive tumors, and these relations affected other clinical parameters, including histology. These analyses stressed the importance of considering joint clinical parameters when assessing hormone effects. The AARP study has also been useful for evaluating effects on cancer risk of other medications. Analyses have also assessed breast cancer risk in relation to non-steroidal anti-inflammatory medications. This area of research has been of interest given that inflammation has been proposed as being involved as an etiologic agent for this cancer site. These analyses showed evidence of reduced risk associated with aspirin use, particularly for estrogen receptor positive cancers. We are also evaluating effects on breast cancer risk of certain medications used in the treatment of cardiovascular diseases that are known to have demethylating effects on DNA. Recent cohort studies demonstrated reduced breast cancer risks among women with a history of fractures or low bone mineral density. The impact of the severity and timing of bone loss on risk has not yet been investigated, and the extent to which other risk factors (family history, anthropometric factors, physical activity, and exogenous hormones) modify the relationship with bone density is unknown. To elaborate on these research questions, we have conducted a follow-up study of over 20,000 postmenopausal women who volunteered for a clinical trial of the bone-enhancing drug alendronate. This large cohort includes extensive baseline information on major breast cancer risk factors, and thus is ideal for evaluating potential interactions with bone mineral density and the effects of bone mineral density on other cancer sites. The availability of serologic samples from study participants will also enable assessment of the interactive effects of endogenous hormones and bone mineral density on subsequent breast cancer risk.